An Adaptive and Privacy-Aware Federated Learning Framework for Efficient and Secure Model Training Across Heterogeneous Datasets

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2026 | Volume : 13 | 01 | Page :
    By

    Pushpendra Kumar Sikarwal,

  • Mukesh Kumar Gupta,

  • Adamya Gupta,

  1. Research Scholar, Department of Computer Science, Suresh Gyan Vihar University, Jaipur, Rajasthan,
  2. Professor, Department of Electrical Engineering, Suresh Gyan Vihar University, Jaipur, Rajasthan, India
  3. Research Scholar, Department of Computer Science and Engineering, Jaipur Engineering College & Research Centre, Jaipur, Rajasthan, India

Abstract

The problem of efficiency and privacy regarding heterogeneous data in modern distributed machine learning systems is a vital point that should be taken into account. The absence of IID data distribution, client heterogeneity, and privacy invasion during the aggregation model are the bane of conventional Federated Learning (FL) approaches to learning like FedAvg and FedProx. The paper proposes that the Adaptive and Privacy-Aware Federated Learning Framework (AFL-P) can be used to address these limitations to ensure that dynamic optimization and hybrid privacy preservation can be achieved. The proposed framework implements adaptive client participation and weighted aggregation with reference on local resource availability and convergence measure thereby improving the efficiency of communication and the model stability. Furthermore, AFL-P is an algorithm that combines Differential Privacy (DP) and Secure Aggregation (SA) to provide a stringent assurance of information leakage and no performance costs on the learning process. The CIFAR-10 (image), UCI-HAR (sensor), and Google Speech Commands (audio) experimental results show that AFL-P outperforms other baseline algorithms (FedAvg, DP-FedAvg and FedProx) by 6–8 percent, 20 percent communication overhead, and more than 50 percent privacy loss. The findings confirm AFL-P is a strong, efficient and privacy- conscientious training model of heterogeneous and resources-constrained set ups.

Keywords: Adaptive optimization, differential privacy (DP), federated learning (FL), heterogeneous data environments, secure aggregation (SA)

How to cite this article:
Pushpendra Kumar Sikarwal, Mukesh Kumar Gupta, Adamya Gupta. An Adaptive and Privacy-Aware Federated Learning Framework for Efficient and Secure Model Training Across Heterogeneous Datasets. Journal of Mobile Computing, Communications & Mobile Networks. 2026; 13(01):-.
How to cite this URL:
Pushpendra Kumar Sikarwal, Mukesh Kumar Gupta, Adamya Gupta. An Adaptive and Privacy-Aware Federated Learning Framework for Efficient and Secure Model Training Across Heterogeneous Datasets. Journal of Mobile Computing, Communications & Mobile Networks. 2026; 13(01):-. Available from: https://journals.stmjournals.com/jomccmn/article=2026/view=240854


References

  1. Brauneck A, Schmalhorst L, Kazemi Majdabadi MM, Bakhtiari M, Völker U, Baumbach J, Baumbach L, Buchholtz G. Federated machine learning, privacy-enhancing technologies, and data protection laws in medical research: scoping review. Journal of medical Internet research. 2023 Mar 30;25:e41588.
  2. Jiang H, Pei J, Yu D, Yu J, Gong B, Cheng X. Applications of differential privacy in social network analysis: A survey. IEEE transactions on knowledge and data engineering. 2021 Apr 13;35(1):108- 27.
  3. Haripriya R, Khare N, Pandey M, Biswas S. A privacy-enhanced framework for collaborative Big Data analysis in healthcare using adaptive federated learning aggregation. Journal of Big Data. 2025 May 6;12(1):113.
  4. Wang B, Feng D, Su J, Song S. An effective federated object detection framework with dynamic differential privacy. Mathematics. 2024 Jul 9;12(14):2150.
  5. Li Y, Yang S, Ren X, Shi L, Zhao C. Multi-stage asynchronous federated learning with adaptive differential privacy. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023 Nov 13;46(2):1243-56.
  6. Zhan S, Huang L, Luo G, Zheng S, Gao Z, Chao HC. A Review on Federated Learning Architectures for Privacy-Preserving AI: Lightweight and Secure Cloud–Edge–End Collaboration. Electronics. 2025 Jun 20;14(13):2512.
  7. Janardhanan H. Federated Learning in Edge Computing: Advancements, Security Challenges, and Optimization Strategies. In2025 8th International Conference on Circuit, Power & Computing Technologies (ICCPCT) 2025 Aug 7 (pp. 1144-1150). IEEE.
  8. Zhou X, Liang W, She J, Yan Z, Wang KI. Two-layer federated learning with heterogeneous model aggregation for 6g supported internet of vehicles. IEEE Transactions on Vehicular Technology. 2021 May 6;70(6):5308-17.
  9. Mohammadi S, Balador A, Sinaei S, Flammini F. Balancing privacy and performance in federated learning: A systematic literature review on methods and metrics. Journal of Parallel and Distributed Computing. 2024 Oct 1;192:104918.
  10. Jog S, Palaniappan D, Jabbar MA. An adaptive framework for privacy-preserving analytics in federated intrusion detection. Decision Analytics Journal. 2025 Sep 24:100641.
Ahead of Print Subscription Original Research
Volume 13
01
Received 01/02/2026
Accepted 06/02/2026
Published 24/04/2026
Publication Time 82 Days
2

Login


My IP

PlumX Metrics